The anion photoelectron (PE) spectra of Ce2Oy (y = 1, 2), Ce3Oy (y = 0–4), Ce4Oy (y = 0–2), and Ce5Oy (y = 1, 2) are reported and analyzed with supporting results from density functional theory calculations. The PE spectra all exhibit an intense electronic transition to the neutral ground state, all falling in the range of 0.7 to 1.1 eV electron binding energy, with polarization dependence consistent with detachment from diffuse Ce 6s-based molecular orbitals. There is no monotonic increase in electron affinity with increasing oxidation. A qualitative picture of how electronic structure evolves with an oxidation state emerges from comparison between the spectra and the computational results. The electronic structure of the smallest metallic cluster observed in this study, Ce3, is similar to the bulk structure in terms of atomic orbital occupancy (4f 5d2 6s). Initial cerium cluster oxidation involves largely ionic bond formation via Ce 5d and O 2p orbital overlap (i.e., larger O 2p contribution), with Ce—O—Ce bridge bonding favored over Ce=O terminal bond formation. With subsequent oxidation, the Ce 5d-based molecular orbitals are depleted of electrons, with the highest occupied orbitals described as diffuse Ce 6s based molecular orbitals. In the y ≤ (x + 1) range of oxidation states, each Ce center has a singly occupied non-bonding 4f orbital. The PE spectrum of Ce3O4 is unique in that it exhibits a single nearly vertical transition. The highly symmetric structure predicted computationally is the same structure determined from Ce3O4+ IR predissociation spectra [A. M. Burow et al., Phys. Chem. Chem. Phys. 13, 19393 (2011)], indicating that this structure is stable in −1, 0, and +1 charge states. Spectra of clusters with x ≥ 3 exhibit considerable continuum signal above the ground state transition; the intensity of the continuum signal decreases with increasing oxidation. This feature is likely the result of numerous quasi-bound anion states or two-electron transitions possible in molecules with abundant nearly degenerate partially occupied orbitals.

Bulk cerium oxides have electronic properties that are desirable for applications in electronics1–3 and catalysis.4–10 It has recently been observed that the more oxidation states a metal can access, the more diverse its applications can be,11 and cerium oxides can famously switch between the +3 and +4 oxidation states with changes in O2 partial pressure (a feature that makes it particularly useful catalytic converters).12–14 Recently, several high profile papers have propelled interest in cerium oxides as a support material for platinum because of the strong support/catalysts interactions that evidently enhance the activity toward the water-gas shift reaction.6,15 Further enhancement is achieved with nanoscale cerium oxide support, which points to the importance of local interactions and defect sites in catalyst activity. Recent theoretical studies have highlighted the role of oxygen vacancies on ceria for coupling water to the Pt (and other) catalysts.15–18 

Our research program has explored metal oxide clusters in a wide range of lower-than traditional oxidation states as models for oxygen-vacancy defect sites,19–21 which are commonly implicated in catalyst activity.22–24 Because of the local nature of defect sites and catalyst-substrate interactions, metal oxide cluster models offer a helpful alternative approach to surface studies.25 Recently, we completed a study on reactions between water and small cerium oxide cluster anions that showed very distinct oxidation state-dependent product distributions.26 Other groups have also used cluster models to study the properties of cerium oxide. For example, Akin et al. found patterns in the photodissociation of CexOy+ clusters with a strong distinction between clusters for which y = 2x − 1 and y > 2x − 1, indicating a change in how O-atoms bind to the cationic clusters at this particular change in stoichiometry. Wu et al. observed size-dependent rates of reactions between CenO2n+1 and CO for CO2 formation.27 Cerium oxyhydroxide cluster anions studied by Aubriet et al. showed the propensity to form species in which the cerium atoms were fully oxidized (+4).28 Hirabayashi et al. reported results of several reactivity studies on cerium suboxide cluster cations showing average Ce-oxidation state dependence on reactions leading to oxidation of CO and NO,29 as well as cerium suboxide cluster cation reactions with O2.30 Mafuné and co-workers have also probed reactions between small CexOy+ suboxide31 and hyperoxide32 clusters with a range of small molecules.33 

Very few spectroscopic studies of cerium oxides have been reported beyond the CeO diatomic, which has been the subject of numerous spectroscopic34–38 and ab initio39–42 investigations, and the insightful ligand field theory (LFT) treatment applied by Field.43 Asmis and co-workers used infrared predissociation spectroscopy combined with DFT calculations to determine the structures of a range of small CexOy+ suboxide clusters, and showed that small clusters bore structural similarities to the bulk.44 They also have similarly determined the structures of small cerium/vanadium oxide cluster cations.45 Willson and Andrews with co-workers measured the matrix-isolated IR spectra of a series of early lanthanide metal containing molecules generated by laser ablation and reactions with small molecules.46–49 We have obtained the anion photoelectron (PE) spectra of CeO and Ce(OH)2,50 and Chi et al. measured low-energy photoelectron images of CeO along with several other LnO diatomics (Ln = La through Nd).51 The anion PE spectroscopic technique offers the advantage of mass specificity and has been applied toward a number of lanthanide metal-based species.52–56 

Here, we present the photoelectron (PE) spectra of a series of CexOy (x = 2–5; y < 2x) clusters in very low oxidation states. The spectra generally exhibit transitions to close-lying neutral electronic states anticipated to arise in electronically complex species. Consider the cerium atom: The 1G4 ground state of the atom arises from the 4f 5d 6s2 orbital occupancy (bulk Ce has 4f 5d2 6s), and ca. 60 electronic states are found within 1 eV of the ground state, a consequence of energetically close-lying 4f, 5d, and 6s orbitals.57 In analyzing the PE spectra with qualitative portraits of the electronic structures of the anion and neutrals species provided from DFT calculations, we map out the evolution of the electronic structure as a function of oxidation, which gives insight into the initial stages of Ce oxidation, a highly exothermic process (Ce powder is pyrophoric). As seen in a previous study on LaOn for n = 1–5,52 the electron affinities of the CexOy clusters do not increase smoothly with oxidation state, as is commonly observed for main group and transition metals.58 This report will be followed by a companion study on the electronic structure of the products formed from reactions between these clusters and water.59 

CexOy PE spectra were collected using an apparatus described in detail previously.60 CexOy cluster anions were generated using a laser ablation/pulsed molecular beam valve source.61 Approximately 3-5 mJ/pulse of the second harmonic (532 nm) output of a Nd:YAG laser, operating at a repetition rate of 30 Hz, was used to ablate the surface of a metal target composed of compressed Ce (Alfa Aesear) powder. The resulting plasma was entrained by a pulse of ultra-high purity helium buffer gas (40 psi backing pressure) introduced from a solenoid-type molecular beam valve, and swept through a 2.5 cm-long, 0.3-cm diameter channel into a vacuum chamber. After collimation of the gas mixture by a 3-mm skimmer, the anions were accelerated on axis into a 1.2-m beam-modulated time-of-flight mass spectrometer.

Before colliding with a dual microchannel plate ion detector, the anions were selectively photodetached using the second (532.1 nm, 2.330 eV) or third (354.7 nm, 3.495 eV) harmonic output of a second Nd:YAG laser at the intersection of the ion drift tube and a perpendicular 1-m field-free drift tube. The drift times of the small fraction of photoelectrons that traveled the length of the drift tube and collided with a second dual microchannel detector assembly were recorded on a digitizing oscilloscope.

Spectra were accumulated between 80 000 and 200 000 laser shots, and were measured with laser polarizations parallel (θ = 0°) and perpendicular (θ = 90°) to the electron drift tube with intensities I0 and I90, respectively, in order to determine the asymmetry parameter, β(E),

βE=I0I9012I0+I90,
(1)

which can be related to the symmetry of the molecular orbital associated with electron detachment.

For calibration purposes, the drift times were converted to electron kinetic energy (eKE) by identifying common transitions observed in spectra of various Ce-based anions with similar electron affinities (EA), collected using both photon energies, and setting the difference in the electron kinetic energies (eKE) to the fundamental energy (1.1650 ± 0.0001 eV), the difference between the energies of the second and third harmonics, using the relationship,

2(1.1650eV)me=t3νto2t2νto2,
(2)

where me is the electron mass and t3ν is the drift time of electrons associated with a selected transition observed in the spectrum obtained using 3.49 eV photon energy that can readily be correlated with a transition in the spectrum obtained with 2.33 eV, appearing at t2ν. The equation is solved for ℓ, and plotted as a function of to. The intersection between this line and other lines generated from several sets of transitions observed in 3.49 eV and 2.33 eV gives a unique ℓ and to, the calibration parameters necessary to compute the eKE values from electron drift times. The eKE values are related to the anion and neutral states via (2) taking into account internal energies for the anion and neutral species,

eKE=hνEATeneutral+Teanion.
(3)

The data presented show electron counts plotted as a function of eBE,

eBE=hνeKE.
(4)

The eBE values reflect the energy difference between the final neutral state and the initial anion state, and are independent of the photon energy used. Laboratory to center-of-mass frame corrections were made to the eKE (and eBE) values.

Molecular and electronic structures of CexOy anions and neutrals were explored using density functional theory. Calculations were performed using the unrestricted B3LYP hybrid method within the Gaussian 09 program suite.62 This method was chosen because of helpful results obtained for CeO, Ce(OH)2, PrO, EuH, EuO, and EuOH anion and neutral species in previous studies.50,55,56 To incorporate relativistic effects on the Ce metal atom, the Stuttgart RSC ANO/ECP basis set with 28 core electrons and contraction of (14s 13p 10d 8f 3g)/[6s 6p 5d 4f 3g] type, developed by Cao and Dolg, was employed,63 with the Dunning-style correlation consistent basis set, aug-cc-pVTZ for the oxygen atoms. Geometry optimization and frequency calculations were performed for all the anion and neutral species in multiple spin states. Numerous initial guess structures in a large range of possible spin states were attempted for all anions and neutrals. The results presented below focus on the lowest energy structures found for the anions (all structural isomers which were found to be energetically competitive were considered) and the neutral species accessible from those anions. The supplementary material includes figures presenting all structures that converged for the anions and neutrals, and their relative energies. Generally, we found structures with more Ce—O—Ce bridge bonds to be more stable than those with Ce=O terminal bonds. Adiabatic detachment energies (ADEs) of 1 − e transitions between anion and neutral states with comparable structures were calculated from the difference between the zero point-corrected energies of the optimized anion and neutral species structures. Photodetachment spectroscopic parameters were gleaned from the optimized anion and neutral structures, vibrational frequencies, and normal coordinates and used to generate simulated spectra using home-written LabView codes for a more quantitative comparison between the experimental and computational results.

A characteristic mass spectrum of the CexOy anion distribution generated by ablation of Ce metal is shown in Figure 1. For each value of x, there is a fairly narrow range of suboxide species, with the CexO(y = 1) cluster being most abundant for all values of x. In all cases, with the exception of x = 2, the metallic cluster (y = 0) is observed, though Ce5 was not observed in sufficient quantities to measure the spectrum. PE spectra and computational results are summarized in Figures 2–14. A general observation on the computational results is that the structures of most species show minor distortion from high symmetry structures, and the orbitals are often highly spin polarized, making electronic state symmetry assignments difficult. Additionally, for all the CexOy/CexOy clusters, there are 7x nearly degenerate 4f-based molecular orbitals, and different occupancies will result in different overall electronic term symmetry. The more salient features are the qualitative orbital descriptions and spin states. In cases where the overall symmetry is ambiguous because of distortion and polarization, the states are simply labeled 2S+1A.

FIG. 1.

Mass spectrum of CexOy cluster distribution using the laser ablation source.

FIG. 1.

Mass spectrum of CexOy cluster distribution using the laser ablation source.

Close modal
FIG. 2.

PE spectra of (a) Ce2O and (b) Ce2O2 collected using 2.330 and 3.495 eV photon energies with laser polarizations parallel (dark green and blue traces) and perpendicular (light green and blue traces) to the direction of electron detection. Features indicated with an asterisk (*) are tentatively assigned to shake-up transitions.

FIG. 2.

PE spectra of (a) Ce2O and (b) Ce2O2 collected using 2.330 and 3.495 eV photon energies with laser polarizations parallel (dark green and blue traces) and perpendicular (light green and blue traces) to the direction of electron detection. Features indicated with an asterisk (*) are tentatively assigned to shake-up transitions.

Close modal
FIG. 3.

Lowest energy Ce2O anion and neutral structures and orbital occupancies. More comprehensive details on relative energies of these and higher energy structures in different spin states are included in the supplementary material. The molecular orbital depictions are labeled with orbital energies relative to the anion HOMO.

FIG. 3.

Lowest energy Ce2O anion and neutral structures and orbital occupancies. More comprehensive details on relative energies of these and higher energy structures in different spin states are included in the supplementary material. The molecular orbital depictions are labeled with orbital energies relative to the anion HOMO.

Close modal
FIG. 4.

Simulations (blue and red traces) of the Ce2O PE spectra (black trace). The two simulations are color-coordinated with the orbital occupancy shown in Fig. 3, with the [4fa4fb]σ6s2(π5d)σ6s*2 initial anion state.

FIG. 4.

Simulations (blue and red traces) of the Ce2O PE spectra (black trace). The two simulations are color-coordinated with the orbital occupancy shown in Fig. 3, with the [4fa4fb]σ6s2(π5d)σ6s*2 initial anion state.

Close modal
FIG. 5.

The lowest energy anion and neutral structures and electronic states calculated for Ce2O2 and orbital occupancies. More comprehensive details on relative energies of different structures in different spin states are included in the supplementary material. The molecular orbital depictions are labeled with orbital energies relative to the anion HOMO.

FIG. 5.

The lowest energy anion and neutral structures and electronic states calculated for Ce2O2 and orbital occupancies. More comprehensive details on relative energies of different structures in different spin states are included in the supplementary material. The molecular orbital depictions are labeled with orbital energies relative to the anion HOMO.

Close modal
FIG. 6.

Simulations (solid blue, dotted blue, and red traces) of the Ce2O2 PE spectrum (black trace). The simulations are color-coordinated with the orbital occupancy shown in Figure 5.

FIG. 6.

Simulations (solid blue, dotted blue, and red traces) of the Ce2O2 PE spectrum (black trace). The simulations are color-coordinated with the orbital occupancy shown in Figure 5.

Close modal
FIG. 7.

PE spectra of (a) Ce3 and (b) Ce3O collected using 2.330 and 3.495 eV photon energies with laser polarizations parallel (dark green and blue traces) and perpendicular (light green and blue traces) to the direction of electron detection.

FIG. 7.

PE spectra of (a) Ce3 and (b) Ce3O collected using 2.330 and 3.495 eV photon energies with laser polarizations parallel (dark green and blue traces) and perpendicular (light green and blue traces) to the direction of electron detection.

Close modal
FIG. 8.

PE spectra of (a) Ce3O2, (b) Ce3O3, and (c) Ce3O4 collected using 2.330 and 3.495 eV photon energies with laser polarizations parallel (dark green and blue traces) and perpendicular (light green and blue traces) to the direction of electron detection.

FIG. 8.

PE spectra of (a) Ce3O2, (b) Ce3O3, and (c) Ce3O4 collected using 2.330 and 3.495 eV photon energies with laser polarizations parallel (dark green and blue traces) and perpendicular (light green and blue traces) to the direction of electron detection.

Close modal
FIG. 9.

Representative low-energy anion and neutral structures and orbital occupancies for Ce3 and Ce3O. More comprehensive details on relative energies of different structures in different spin states are included in the supplementary material. The molecular orbital depictions are labeled with orbital energies relative to the anion HOMO.

FIG. 9.

Representative low-energy anion and neutral structures and orbital occupancies for Ce3 and Ce3O. More comprehensive details on relative energies of different structures in different spin states are included in the supplementary material. The molecular orbital depictions are labeled with orbital energies relative to the anion HOMO.

Close modal
FIG. 10.

Representative low-energy anion and neutral structures and orbital occupancies for Ce3 and Ce3O. More comprehensive details on relative energies of different structures in different spin states are included in the supplementary material. The molecular orbital depictions are labeled with orbital energies relative to the anion HOMO.

FIG. 10.

Representative low-energy anion and neutral structures and orbital occupancies for Ce3 and Ce3O. More comprehensive details on relative energies of different structures in different spin states are included in the supplementary material. The molecular orbital depictions are labeled with orbital energies relative to the anion HOMO.

Close modal
FIG. 11.

Competitive anion and neutral structures and term symbols calculated for Ce3O3 and Ce3O4. The orbital layout is labeled with orbital energies relative to the ground state anion to neutral transition.

FIG. 11.

Competitive anion and neutral structures and term symbols calculated for Ce3O3 and Ce3O4. The orbital layout is labeled with orbital energies relative to the ground state anion to neutral transition.

Close modal
FIG. 12.

Simulations (red traces) of the Ce3O3 and Ce3O4 PE spectra (black trace). The colored solid line traces denote the identified transitions found in the orbital diagram [Fig. 11], see text for details. The determined electron detachment orbital is illustrated next to the respective transitions.

FIG. 12.

Simulations (red traces) of the Ce3O3 and Ce3O4 PE spectra (black trace). The colored solid line traces denote the identified transitions found in the orbital diagram [Fig. 11], see text for details. The determined electron detachment orbital is illustrated next to the respective transitions.

Close modal
FIG. 13.

PE spectra of (a) Ce4, (b) Ce4O, and (c) Ce4O2 collected using 2.330 and 3.495 eV photon energies with laser polarizations parallel (dark green and blue traces) and perpendicular (light green and blue traces) to the direction of electron detection.

FIG. 13.

PE spectra of (a) Ce4, (b) Ce4O, and (c) Ce4O2 collected using 2.330 and 3.495 eV photon energies with laser polarizations parallel (dark green and blue traces) and perpendicular (light green and blue traces) to the direction of electron detection.

Close modal
FIG. 14.

PE spectra of (a) Ce5O and (b) Ce5O2 collected using 2.330 and 3.495 eV photon energies with laser polarizations parallel (dark green and blue traces) and perpendicular (light green and blue traces) to the direction of electron detection.

FIG. 14.

PE spectra of (a) Ce5O and (b) Ce5O2 collected using 2.330 and 3.495 eV photon energies with laser polarizations parallel (dark green and blue traces) and perpendicular (light green and blue traces) to the direction of electron detection.

Close modal

PE spectra of Ce2O and Ce2O2 are shown in Figures 2(a) and 2(b), respectively, with transition energies summarized in Table I. The PE spectra were obtained using 3.495 eV (blue traces) and 2.330 eV (green traces) photon energies. The dark blue and green traces are spectra collected with the laser polarization parallel to the photoelectron drift tube (θ = 0°), and the light green and blue traces are spectra collected with perpendicular polarization (θ = 90°).

TABLE I.

Peak positions and tentative assignments for the PE spectra of Ce2O and Ce2O2 [Fig. 2]. The values in parentheses represent the uncertainty in the last digit.

ADE/VDE eBE/eVAsymmetry parameterTentative assignment
Ce2O    
1.06(2)/1.09(2) 0.5(1) 5A14B2 or 5B14A2 
VDE = 1.16(2) 0.8(1) 3A14B2 or 3B14A2 
VDE = 2.3(1) … Shake-up transition associated with promotion of 6s-based MO to 5d-based MO 
Ce2O2    
0.83(2)/0.88(1) 1.2(1) 3B1u4Ag 
1.05(2)/1.09(1) 0.9(1) 5Ag4Ag 
VDE = 2.0(1) … Shake-up transition associated with promotion of 6s-based MO to 5d-based MO 
ADE/VDE eBE/eVAsymmetry parameterTentative assignment
Ce2O    
1.06(2)/1.09(2) 0.5(1) 5A14B2 or 5B14A2 
VDE = 1.16(2) 0.8(1) 3A14B2 or 3B14A2 
VDE = 2.3(1) … Shake-up transition associated with promotion of 6s-based MO to 5d-based MO 
Ce2O2    
0.83(2)/0.88(1) 1.2(1) 3B1u4Ag 
1.05(2)/1.09(1) 0.9(1) 5Ag4Ag 
VDE = 2.0(1) … Shake-up transition associated with promotion of 6s-based MO to 5d-based MO 

The spectrum of Ce2O exhibits two overlapping transitions near an eBE value of 1.1 eV, labeled X and A, respectively. The origin of band X is 1.06(2) eV, which can be taken as the EA of Ce2O. The vertical detachment energy (VDE) of band X is 1.09(2) eV. The origin of band A is difficult to identify because of overlap with band X; the VDE is 1.16(2) eV. The values in parentheses represent the uncertainty in the last digit. At first glance, band A could conceivably be a member of a 560 cm−1 vibrational progression originating from peak X. However, based on the profile of the spectrum and on the computational results (vide infra), we tentatively assign these features to distinct electronic transitions. Bands X and A have asymmetry parameters of 0.5(1) and 0.8(1), respectively.

The spectrum of Ce2O2 [Fig. 2(b)] shows two very distinct bands labeled X and A at 0.83(2) and 1.05(2) eV, respectively. Both bands exhibit partially resolved but irregular structure. Bands X and A have asymmetry parameters of 1.2(1) and 0.9(1), respectively. We note here that the EA of Ce2O2, which can be taken to be 0.83(2) eV, is 0.23 eV lower than the EA of Ce2O.

Both spectra exhibit low intensity features at approximately 1.2 eV higher in binding energy relative to band X, indicated by asterisks (*). Similar features have been observed in the PE spectra of CeO and PrO, and were tentatively assigned to shake-up transitions.50,55 For both CeO and PrO, the HOMO is a doubly occupied 6s-like σ orbital. The energetically viable shake-up transitions involve detachment of one of the 6s electrons accompanied by promotion of the other 6s electron into a 5d-like orbital. CeO and PrO neutral excited states that arise from the 4fn5d superconfigurations approximately 1 eV higher in energy than the 4fn6s ground state superconfiguration. Based on the calculations presented in the next several paragraphs, the transitions observed in the Ce2O and Ce2O2 also involve detachment from molecular orbitals that have large contributions from Ce 6s orbitals, so we similarly tentatively assign the features labelled (*) to shake-up transitions.

Results of calculations on Ce2O anions and neutrals suggest that Ce—O—Ce bridge bonding is a favored structural motif, as summarized in Figure 3. Table II includes results on several of the anion and neutral states to illustrate how close-lying the various states and the predicted transition energies are. Atomic coordinates for these and additional states along with more details from the computational outputs are included in the supplementary material. C2v structures were clearly favored for Ce2O, while a nearly linear but low-symmetry Ce—O—Ce (∠Ce—O—Ce = 171°) structure was calculated to be the lowest energy neutral structure, with C2v structures found 0.15 eV higher in energy.

TABLE II.

Summary of results of DFT calculations on the molecular and electronic structures of Ce2O/Ce2O and Ce2O2/Ce2O2 including bond lengths and vibrational frequencies. A more extensive catalog of computational results is included in the supplementary material.

Electronic stateRelative energy (eV)Initial state for 1 − e allowed transition; energy (eV)Vibrational frequency (cm−1)Ce—Ce//Ce—O distance (Å)Expt. EA (eV)
Ce2  Bend/asym str/sym str   
3B1(π5d1.11 4A2; 1.06 94/363/463 3.511//2.064 1.06(2) 
3A1(σ5d1.09 2B2/4B2; 0.99/1.09 107/362/489 3.397//2.061 
5B1(π5d0.99 4A2; 0.94 95/244/474 3.434//2.058 
5A1(σ5d0.93 4B2; 0.93 99/295/460 3.500//2.065 
3A″ 0.78 … 29/509/203 4.107//2.072/2.06 
Ce2O   Bend/asym str/sym str   
2B2(σ5d0.11  80/399/425 3.642/2.064  
4A2(π5d0.05  42/262/377 3.779//2.070  
4B2(σ5d 67/273/406 3.701//2.068  
Ce2O2   Sym. bend/sym str   
5Ag 0.93 4Ag; 0.93 228/586 3.278//2.101 0.83(2) 
3B1u 0.85 4Ag/2Ag; 0.85/0.83 231/583 3.220//2.087 
Ce2O2   Sym. bend/sym str   
2Ag 0.02  224/573 3.262//2.104  
4Ag  225/573 3.262//2.104  
Electronic stateRelative energy (eV)Initial state for 1 − e allowed transition; energy (eV)Vibrational frequency (cm−1)Ce—Ce//Ce—O distance (Å)Expt. EA (eV)
Ce2  Bend/asym str/sym str   
3B1(π5d1.11 4A2; 1.06 94/363/463 3.511//2.064 1.06(2) 
3A1(σ5d1.09 2B2/4B2; 0.99/1.09 107/362/489 3.397//2.061 
5B1(π5d0.99 4A2; 0.94 95/244/474 3.434//2.058 
5A1(σ5d0.93 4B2; 0.93 99/295/460 3.500//2.065 
3A″ 0.78 … 29/509/203 4.107//2.072/2.06 
Ce2O   Bend/asym str/sym str   
2B2(σ5d0.11  80/399/425 3.642/2.064  
4A2(π5d0.05  42/262/377 3.779//2.070  
4B2(σ5d 67/273/406 3.701//2.068  
Ce2O2   Sym. bend/sym str   
5Ag 0.93 4Ag; 0.93 228/586 3.278//2.101 0.83(2) 
3B1u 0.85 4Ag/2Ag; 0.85/0.83 231/583 3.220//2.087 
Ce2O2   Sym. bend/sym str   
2Ag 0.02  224/573 3.262//2.104  
4Ag  225/573 3.262//2.104  

Again, we do not presume to get an exact description of the electronic structure of these clusters from the computational results, but in qualitative terms, the electronic structure of the [CeaOCeb] anions can be described as ...[4fa4fb]σ6s2[CeaCebbondingorbital](5d)(σ6s*)2. In this description, we have foregone the symmetry of the orbitals in favor of a description of the bonding character of the orbitals. The 4f-based molecular orbitals have very little overlap, and their occupancy can be described as a single electron occupying a single 4f orbital on each of the two Ce atoms. These 4f orbitals lie energetically between the Ce—O bonding orbitals and the highest-lying Ce 5d and 6s-based molecular orbitals. The Ce—O bonding orbitals are dominated by O 2p orbitals, with some contribution from Ce 5d orbitals, consistent with predominantly ionic bonding. The term symbols included in Fig. 3 take into account the symmetry of the 4f orbitals, with the caveat that shifting electrons from the orbitals depicted here into any of the other 4f orbitals is very unlikely to change the overall term energy significantly, since from a bonding standpoint, these orbitals are core-like.

The exact identity of the singly occupied Ce 5d–Ce 5d based orbital also does not affect the overall bonding picture in Ce2O. Based on the calculations, occupation of either a σ5d or π5d orbital, both of which are shown in Fig. 3, gives nearly isoenergetic states, with the σ5d occupancy favored by 0.06 eV. States with doublet spin multiplicity formed when the spin of the electron in the 5d-5d orbital is antiparallel to the two unpaired electrons in the 4f orbitals are predicted to be 0.1–0.2 eV higher in energy than the quartet states. The 5d-based orbitals are also nearly isoenergetic with, though slightly lower in energy than, the σ6s* orbital, so numerous neutral states predicted to be within 0.3 eV of the calculated ground neutral state could in principle be accessed via detachment of numerous close-lying anion states.

As a starting point for assigning the spectrum, we generated simulations based on computational results on one of the nearly degenerate lowest-energy quartet states of the anion, [4fa4fb]σ6s2(π5d)(σ6s*)2 to the quintet (“blue” electron) and triplet states (“red” electron) with [4fa4fb]σ6s2(π5d)(σ6s*) orbital occupancy, shown in Figure 4. Simulations based on results for the analogous quintet and triplet states with the singly occupied σ5d orbital rather than π5d were nearly identical. Vibrational frequencies and general structural parameters are summarized in Table II; a more extensive list of spectroscopic parameters (normal coordinate displacements, transition origins, peak widths, and temperature) is included in the supplementary material. The 5B14A2 transition (blue trace) features fairly extended progressions in the 95-100 cm−1 bend mode (2 cm−1 anharmonicity was used in calculating peak positions), which is unresolved in our spectrum, and very little excitation of the 460–470 cm−1 symmetric stretch. The symmetric stretch progression is slightly more extended in the 3B14A2 simulation. Detachment of an electron from the σ6s* orbital affects the Ce—Ce equilibrium bond distance, as summarized in Table II, but is fairly non-bonding with respect to Ce—O.

The transition from either of the quartet anion states to the nearly linear ...[4fa4fb]σ6s2(σ6s*)2 state (detachment of the “green” electron, Fig. 3) is calculated to originate at lower electron binding energy, but based on the substantial structural change associated with this transition as well as the smaller photodetachment cross sections for 5d versus 6s orbital detachment,64 we tentatively assign the very low-intensity signal observed from ca. 0.7 eV up to the origin of band X to this transition. Note that there is also broad signal to higher binding energy in PE spectrum. Because the σ6s and σ6s* orbitals are separated by less than 1 eV in the calculations, it is possible that detachment from the σ6s orbital may contribute to some of the signals in the eBE > 1.3 eV range.

Figure 5 shows the lowest energy structure calculated for Ce2O2, which is nearly identical to the lowest energy structure found in calculations on Ce2O2, both with D2h symmetry. Depictions of the highest singly and doubly occupied orbitals are also shown. In broad terms, the electronic structure of Ce2O2 can be related to the Ce2O molecule as having Ce—O bonding orbitals that are dominated by O 2p orbitals (three of the six are shown in Fig. 5), one singly occupied 4f orbital on each of the two Ce atoms, and as expected, two fewer electrons in the higher lying Ce-local orbitals, which can be described as σ6s and σ6s*, with an overall occupancy of ...[4fa4fb](σ6s)2(σ6s*). The anion is predicted to have a 4Ag ground state, with the 2Ag state in which the σ6s* electron spin is antiparallel to the two 4f electrons 0.02 eV higher in energy. This very small quartet-doublet splitting is comparable to an analogous splitting in the CeO diatomic molecule: the X12(jf = 5/2, js = − 1/2) − X13(jf = 5/2, js = 1/2) splitting, 0.01 eV, reflects how decoupled electrons in the 4f and 6s orbitals are. The predicted 3B1u neutral ground state has the ...[4fa 4fb](σ6s)2 orbital occupancy, and the 5Ag state with ...[4fa4fb](σ6s)(σ6s*) occupancy is predicted to be 0.08 eV higher in energy. Calculations on the 3Ag state with the same occupancy but antiparallel spins of the electrons in the (σ6s) and (σ6s*) orbitals did not converge. A summary of the computational results is included in Table II. Again, we note that shifting the Ce 4f-local electrons into alternative nearly degenerate orbitals will result in nearly identical, close-lying states with different overall electronic species.

Based on the overall orbital occupancy of the Ce2O2 anion, we assign band X to transitions associated with detaching an electron from the singly occupied (σ6s*) orbital, and band A to detachment from the (σ6s) orbital. Simulations of the 3B1u4Ag (blue trace) 3B1u2Ag (blue dotted trace) and 5Ag4Ag (red trace) transitions based on calculated spectroscopic parameters are shown in Figure 6. Vibrational frequencies of the totally symmetric modes and general structural parameters are summarized in Table II; a more extensive list of spectroscopic parameters (normal coordinate displacements, transition origins, peak widths, and temperature) is included in the supplementary material. The energies of the origins were shifted by less than 0.1 eV in order to match the spectrum, and the order of the transition energies was preserved. Note that the electronic configurations of the two anion and two neutral states involved in the simulations point to additional transitions, 3Ag2Ag and 1Ag2Ag, slightly higher in energy than the 5Ag4Ag transition, and band A in the experimental spectrum may appear broader than the simulated spectrum because of these overlapping transitions. The near vertical appearance of simulated spectra indicates that the (σ6s) and (σ6s*) orbitals are non-bonding. However, the highest ag stretch frequency determined from calculations, and indicated on Fig. 6, does appear to be active in band A in the experimental spectrum, though as noted above, it is also possible that several electronic transitions are contributing to band A.

The PE spectra of Ce3Oy (y = 0, 1) are shown in Figure 7, and the spectra of Ce3Oy (y = 2–4) are shown in Figure 8. As before, the blue traces are spectra obtained with 3.495 eV photon energy, and the green traces are the spectra obtained using 2.330 eV photon energy; darker traces indicate laser polarization parallel (θ = 0°) to the direction of electron drift, and lighter traces indicate orthogonal (θ = 90°) polarization. The EA’s of this series of clusters again do not change monotonically with oxidation state: In order of increasing oxidation for Ce3 to Ce3O4, they are 0.67(1), 0.89(5), 0.95(5), 1.00(1), and 0.792(5) eV. The ADE and VDE values are summarized in Table III.

TABLE III.

Peak positions and tentative assignments for the PE spectra of Ce3Oy shown in Figs. 7 and 8. The values in parentheses represent the uncertainty in the last digit. Asymmetry parameters for bands that have low signal to noise and/or are embedded in continuum signal were not calculated. No tentative assignments are specified for Ce3, Ce3O, or Ce3O2 because numerous calculated energies of transitions between different structures and spin states of nearly isoenergetic anions are in reasonable agreement with the spectra (see Table IV).

ADE/VDE (eV)Asymmetry parameterTentative assignment
Ce3    
0.67(1)/0.74(1) 1.1(2)  
1.12(2)/1.20(1) 0.2(1)  
VDE = 1.7(1) …  
Ce3O    
0.89(5)/0.91(2) 0.5(1)  
1.0(2)/1.09(1) 0.3(2)  
VDE = 1.5(1) …  
Ce3O2    
0.95(5)/1.10(2) 0.8(1)  
Ce3O3    
1.00(1)/1.10(1) 1.0(1) 5A + e6
VDE = 1.72(5) … Shake-up transition associated with promotion of 6s-based MO to 5d-based MO 
Ce3O4    
0.792(5)/0.792(5) 1.5(2) 5A/3A + e4
VDE = 1.25(3) … Shake-up transition associated with promotion of 6s-based MO to 5d-based MO 
ADE/VDE (eV)Asymmetry parameterTentative assignment
Ce3    
0.67(1)/0.74(1) 1.1(2)  
1.12(2)/1.20(1) 0.2(1)  
VDE = 1.7(1) …  
Ce3O    
0.89(5)/0.91(2) 0.5(1)  
1.0(2)/1.09(1) 0.3(2)  
VDE = 1.5(1) …  
Ce3O2    
0.95(5)/1.10(2) 0.8(1)  
Ce3O3    
1.00(1)/1.10(1) 1.0(1) 5A + e6
VDE = 1.72(5) … Shake-up transition associated with promotion of 6s-based MO to 5d-based MO 
Ce3O4    
0.792(5)/0.792(5) 1.5(2) 5A/3A + e4
VDE = 1.25(3) … Shake-up transition associated with promotion of 6s-based MO to 5d-based MO 

The spectra of Ce3 and Ce3O [Figs. 7(a) and 7(b)] exhibit a dominant band labeled X, with notably higher intensity at θ = 0° laser polarization. Qualitatively, the PE spectra of Ce3 and Ce3O are similar: Band X is fairly narrow, ca. 0.09 eV full width at half maximum (FWHM), and there are two additional broad electronic bands labeled A and B at higher binding energy. Band A in both has a more isotropic angular distribution, with the asymmetry parameter, β(E), approaching zero in the spectrum of Ce3O obtained with 3.49 eV photon energy. Band B in both spectra has a lower intensity and suffers from low signal to noise and accompanying continuum signal, but the polarization dependences for both Ce3 and Ce3O are similar to band X. Relative energies of all bands in the Ce3Oy spectra are included in Table III.

The PE spectra of Ce3Oy(y = 2–4) [Figs. 8(a)8(c)] are also dominated by their respective lowest eBE transition, labeled band X, and all three have varying undulations of lower-intensity signal to higher eBE. As with the spectra of Ce3 and Ce3O, there is continuum signal in the Ce3O2 and Ce3O3 spectra from band X up to eBE = hv, and it is more intense in the Ce3O2 spectrum. Bands X in the Ce3O2 and Ce3O3 spectra are relatively broad, 0.25 eV FWHM. The Ce3O2 spectrum exhibits a progression of shoulders spaced by 500 cm−1, which is typical of a Ce—O—Ce stretch frequency, and can be modeled with a Franck-Condon simulation (not shown) assuming a normal coordinate displacement of ΔQ = 0.75 Å amu1/2. Computational results described below do not point to a clear structural determination, however. The PE spectrum of Ce3O3 has partially resolved vibrational structure that can be reconciled with computational results described below, and a low intensity band labeled with an asterisk (*) at approximately 0.7 eV higher in energy. Band X in the PE spectrum of Ce3O4 is nearly vertical and exhibits a low-intensity and fairly narrow excited state transition labeled with an asterisk (*) at eBE = 1.25(3) eV [T0 = 0.46(3) eV]. Only the spectra obtained with θ = 0° laser polarization are shown; spectra collected with θ = 90°had low signal intensity, giving β(E) ≈ 1.5(2).

Pictorial summaries of the results of calculations on the molecular and electronic structures for the lowest energy anion and neutral trimetallic cerium oxide clusters are shown in Figures 9–11, with a summary of the various electronic state relative energies and several structural and vibrational details found in Table IV. A more extensive listing of computational results is found in the supplementary material.

TABLE IV.

Summary of the lowest lying electronic structures and states predicted from DFT calculations on Ce3Oy (x = 0–4). In several cases, trivial distortions from high-symmetry structures or pronounced spin polarization resulted in ambiguous symmetry determination in the computational output files, though symmetry is evident from visual inspection. These cases are indicated with an asterisk (*). Note that several neutral states determined computationally are not one-electron accessible via photodetachment of any of the low lying anions.

Electronic statePoint groupRelative energy (eV)Initial state and 1 – e allowed transition energy (eV)Ce—Ce//Ce—O distances (Å)Expt. EA (eV)
Ce3      
5A″ C2v 0.96 4A/6A2; 0.89/0.96 3.236/3.021/3.022 0.67(1) 
1A1 C2v 0.95 2B2; 0.85 3.296/2.582/2.582 
1C2v 0.93 2B2; 0.88 2.933/3.064/3.072 
7B1 C2v 0.89 6A2; 0.89 eV 2.867/2.947/2.947 
Ce3      
2B2 C2v 0.10  3.225/3.225/3.150  
4D3h 0.07  3.196/3.196/3.180  
8B2 C2v 0.05  2.911/2.911/3.807  
6A2 C2v  2.724/2.724/3.751  
Ce3     
3C3v (pyr) 0.98 4E; 0.98 3.371/3.223/3.224//2.384/2.185 0.89(5) 
5C3v (pyr) 0.89 4E/6 A; 0.89/0.83 3.189/3.193/3.324//2.361/2.195 
7C3v (pyr) 0.88 6A/0.82 3.261/3.260/2.936//2.350/2.100 
5C1 (kite) 0.68 4A/6A; 0.66/0.63 3.029/3.213/3.288//2.121/2.057 
Ce3O      
6C3v (pyr) 0.06  3.593/3.593/3.250//2.284/2.127  
6C1 (kite) 0.05  3.430/3.309/2.980//2.152/2.092  
2C3v (pyr) 0.05  3.282/3.282/3.282//2.232/2.254  
4C1 (kite) 0.02  3.419/3.327/3.295//2.080/2.802  
2C1 (kite) 0.01  3.354/3.398/3.197//2.064/2.106  
4C3v (pyr)  3.271/3.271/3.271//2.249  
Ce3O2      
3Cs (house) 1.01 2A″/4A″; 1.01/1.00 3.732/3.732/3.742//2.027/2.131 0.95(5) 
3C1 (kite) 0.99 4A; 0.98 3.425/3.643/3.649//2.300/2.084 
5B1 C2v (house) 0.98 4A″; 0.97 3.765/3.765/3.266//2.110/2.110 
5C1 (kite) 0.97 4A/6A; 0.96/0.96 3.425/3.616/3.618//2.300/2.084 
1C1 (kite) 0.97 … 3.654/3.641/3.427//2.300/2.086 
Ce3O2      
4C1 (kite) 0.02  3.659/3.409/4.292//2.082/2.321  
4A″ Cs (house) 0.01  3.783/3.787/4.206//2.002/2.154  
6C1 (kite) 0.01  3.407/3.666/4.287//2.321/2.082  
2A″ Cs (house)  3.780/3.780/4.190//2.006/2.151  
Ce3O3      
3A′ Cs (book) 0.97 4A′; 0.92 3.405/3.405/4.040//2.153/2.343 1.00(1) 
5A′ Cs (book) 0.94 4A′/6A′; 0.89/0.94 3.397/3.397/3.954//2.142/2.324 
1A″ Cs (book) 0.78 … 3.848/3.848/3.555//2.107/2.096 
5A″ Cs (ring) 0.78 … 3.847/3.847/3.554//2.108/2.095 
3B2 C2v (ring) 0.73 … 3.880/3.880/3.933//2.102/2.098 
Ce3O3      
4A′ Cs (book) 0.05  3.451/3.451/3.889//2.182/2.420  
6A′ Cs (book)  3.451/3.451/3.924//2.191/2.045  
Ce3O4      
3C3v 0.57 2A/4 A; 0.38/0.47 3.417/3.417/3.417//2.129/2.282  
5C3v 0.56 4A/6 A; 0.56/0.46 3.416/3.416/3.419//2.131/2.276 0.792(5) 
Ce3O4      
2C3v 0.19  3.428/3.427/3.407//2.129/2.269  
6C3v 0.10  3.417/3.417/3.447//2.141/2.309  
4C3v  3.426/3.426/3.413//2.129/2.307  
Electronic statePoint groupRelative energy (eV)Initial state and 1 – e allowed transition energy (eV)Ce—Ce//Ce—O distances (Å)Expt. EA (eV)
Ce3      
5A″ C2v 0.96 4A/6A2; 0.89/0.96 3.236/3.021/3.022 0.67(1) 
1A1 C2v 0.95 2B2; 0.85 3.296/2.582/2.582 
1C2v 0.93 2B2; 0.88 2.933/3.064/3.072 
7B1 C2v 0.89 6A2; 0.89 eV 2.867/2.947/2.947 
Ce3      
2B2 C2v 0.10  3.225/3.225/3.150  
4D3h 0.07  3.196/3.196/3.180  
8B2 C2v 0.05  2.911/2.911/3.807  
6A2 C2v  2.724/2.724/3.751  
Ce3     
3C3v (pyr) 0.98 4E; 0.98 3.371/3.223/3.224//2.384/2.185 0.89(5) 
5C3v (pyr) 0.89 4E/6 A; 0.89/0.83 3.189/3.193/3.324//2.361/2.195 
7C3v (pyr) 0.88 6A/0.82 3.261/3.260/2.936//2.350/2.100 
5C1 (kite) 0.68 4A/6A; 0.66/0.63 3.029/3.213/3.288//2.121/2.057 
Ce3O      
6C3v (pyr) 0.06  3.593/3.593/3.250//2.284/2.127  
6C1 (kite) 0.05  3.430/3.309/2.980//2.152/2.092  
2C3v (pyr) 0.05  3.282/3.282/3.282//2.232/2.254  
4C1 (kite) 0.02  3.419/3.327/3.295//2.080/2.802  
2C1 (kite) 0.01  3.354/3.398/3.197//2.064/2.106  
4C3v (pyr)  3.271/3.271/3.271//2.249  
Ce3O2      
3Cs (house) 1.01 2A″/4A″; 1.01/1.00 3.732/3.732/3.742//2.027/2.131 0.95(5) 
3C1 (kite) 0.99 4A; 0.98 3.425/3.643/3.649//2.300/2.084 
5B1 C2v (house) 0.98 4A″; 0.97 3.765/3.765/3.266//2.110/2.110 
5C1 (kite) 0.97 4A/6A; 0.96/0.96 3.425/3.616/3.618//2.300/2.084 
1C1 (kite) 0.97 … 3.654/3.641/3.427//2.300/2.086 
Ce3O2      
4C1 (kite) 0.02  3.659/3.409/4.292//2.082/2.321  
4A″ Cs (house) 0.01  3.783/3.787/4.206//2.002/2.154  
6C1 (kite) 0.01  3.407/3.666/4.287//2.321/2.082  
2A″ Cs (house)  3.780/3.780/4.190//2.006/2.151  
Ce3O3      
3A′ Cs (book) 0.97 4A′; 0.92 3.405/3.405/4.040//2.153/2.343 1.00(1) 
5A′ Cs (book) 0.94 4A′/6A′; 0.89/0.94 3.397/3.397/3.954//2.142/2.324 
1A″ Cs (book) 0.78 … 3.848/3.848/3.555//2.107/2.096 
5A″ Cs (ring) 0.78 … 3.847/3.847/3.554//2.108/2.095 
3B2 C2v (ring) 0.73 … 3.880/3.880/3.933//2.102/2.098 
Ce3O3      
4A′ Cs (book) 0.05  3.451/3.451/3.889//2.182/2.420  
6A′ Cs (book)  3.451/3.451/3.924//2.191/2.045  
Ce3O4      
3C3v 0.57 2A/4 A; 0.38/0.47 3.417/3.417/3.417//2.129/2.282  
5C3v 0.56 4A/6 A; 0.56/0.46 3.416/3.416/3.419//2.131/2.276 0.792(5) 
Ce3O4      
2C3v 0.19  3.428/3.427/3.407//2.129/2.269  
6C3v 0.10  3.417/3.417/3.447//2.141/2.309  
4C3v  3.426/3.426/3.413//2.129/2.307  

The anions and neutrals calculated for most species are compact, and all O-atoms are in bridging positions. Results for Ce3/Ce3 and Ce3O/Ce3O are shown in Fig. 9. Again, gleaning a qualitative picture of electronic structures of the anions and neutrals, the Ce atoms in all anions and neutrals can be described as having a singly occupied 4f orbital. The Ce3/Ce3 structures are either D3h or Jahn-Teller distorted C2v structures. The HOMO in the C2v structure is a 6s-6s antibonding orbital with respect to two of the three Ce centers. This orbital is singly occupied in the lowest energy neutral structure and doubly occupied in the anion. Taken together with the a1 (or a1′ for the D3h structure) orbital that can be described as predominantly the in-phase combination of all three 6s orbitals, the HOMO can be interpreted as one of the Jahn-Teller distorted components of the degenerate 6s-based e′ orbitals. Interleaved energetically with the largely 6s-based orbitals are 5d-based bonding orbitals, two of which were predicted to be singly occupied. Neutral Ce3 therefore can be described as three Ce-centers with 4f 5d2 6s occupancy, which is bulk-like.65 

Upon bonding with a single O-atom, the Ce centers in Ce3O are still electron rich. Two nearly isoenergetic C3v and Cs structures in which the oxygen atom is either bound to all three cerium atoms (pyramid structure) or bridging two of the Ce atoms (kite structure) were found computationally (the computational results generally have these structures distorted slightly from C3v and Cs). From a qualitative standpoint, the O-atom bonds with the Ce3 portion again through 2p–5d overlap, though there is more contribution from 2p orbitals, which is consistent with a largely ionic bond. However, comparing the orbital occupancies of Ce3/Ce3 and Ce3O/Ce3O, electron density into the O2− is drawn from the antibonding 6s-based orbital and one of the 5d-based orbitals. The orbitals and occupancies depicted in Fig. 9 correspond to the 4E Ce3O pyramid structure, which was predicted to be marginally lower than the kite structure; on the neutral surface, the kite structure is slightly lower in energy. The calculations predict that the single electron in the Ce3O HOMO is antiferromagnetically (AF) coupled to the 4f electrons, though spin states in which there is some AF coupling between the 4f electrons are also calculated to lie in a very small window of energy.

A pictorial summary of results of calculations on Ce3O2 is shown in Fig. 10. Two structures, house and distorted kite, were found to be very close in energy for both the anion and neutral. The orbitals and occupancies shown are for the distorted kite structure. Comparing the Ce-local orbitals in Ce3O2/Ce3O2 to Ce-local orbitals in Ce3O/Ce3O in Fig. 9, the additional O-atom again bonds primarily through overlap with the 5d-based MO’s, and because of largely ionic character, the resulting bonding orbitals are predominantly 2p-like. Therefore, the 6s-based MO occupancies of Ce3O and Ce3O2 appear similar, with the difference being the occupancy of the 5d-based MO’s.

For both Ce3O and Ce3O2, we cannot determine unambiguously which structures are contributing to the spectra based on the computational results, though from an electronic structure and transition energy standpoint, the different structural isomers are similar.

Calculations on Ce3O3 and Ce3O4 are summarized pictorially in Fig. 11. A Cs book structure was the only low-energy structure found for Ce3O3. Neutral Ce3O3 favors an open ring structure, with the book isomer lying 0.2 eV higher in energy. The Ce centers in Ce3O3 appear to have the same 4f6s orbital occupancy as CeO, with the extra charge in the anion localized in a 6s-like orbital (the quartet and sextet spin states for the anion have the same occupancy, with the unpaired electrons in the highest 6s-based MO’s being either parallel or antiparallel to the electrons in the 6s-based a″ orbital).

As summarized in Table IV, the ADE values calculated for all of these species are in reasonable agreement with the spectra, and it appears that sequential oxidation of Ce3 does not affect the electron affinity significantly. The direct conclusion of this observation is that the anion and neutral cerium clusters are stabilized equally by oxidation. The various close-lying structural isomers found for Ce3O and Ce3O2 also have nearly identical ADE values. Spectral simulations generated using parameters from the unique structures found for Ce3O3 and Ce3O4 are shown in Figures 12(a) and 12(b), respectively. For Ce3O3, we assumed that the unique book structure found for the anion only accesses the neutral book structure (the higher-lying neutral isomer), and the simulation shows only the sextet to quintet transition. The simulated spectrum is better resolved than the experimental spectrum, but we note again that there are close lying electronic states with different spin multiplicities, all of which can be contributing to the spectrum. The origin of the simulation was shifted up by 0.08 eV relative to the calculated ADE to match the experimental profile. Overall, the profile of the simulated and experimental spectra is in reasonable agreement.

The calculation-based simulation of the Ce3O4 spectrum is in near-perfect agreement with the experimental spectrum, though the origin was shifted up by 0.23 eV relative to the calculated ADE value to match the observed spectrum. This disparity is within the ±0.3 eV error in ADE expected in our calculations,66 though it is larger than other results presented here. The spin densities for the anion and neutral states are included in Fig. 12(b). The orbital occupancy shown for the quartet spin state of Ce3O4 [Fig. 11] exhibits broken symmetry, and the two singly occupied (antiparallel spin) orbitals may be a doubly occupied orbital which is expressed with significant spin polarization in the calculations. In the spin-pure neutral quintet state, the singly occupied HOMO is the symmetric combination of the three 6s orbitals.

Qualitatively, we can compare the orbital occupancies and relative energies with the decrease in the number of distinct excited state transitions observed in the spectra as Ce3Oy oxidation is increased. First, the O-local bonding orbitals are energetically inaccessible with the photon energies used in this experiment, generally lying >5 eV lower in energy than the HOMO’s. The 4f-based MOs are also energetically near the upper limit of the photon energy and are core-like non-bonding orbitals with very low photodetachment cross section.64 Therefore, the detachment signal we observe is associated with 5d- and 6s-based orbitals, which are energetically interleaved for Ce3Oy, y = 0–2. The 5d-based orbitals are vacant, for y = 3, 4, and are the anti-bonding counterparts of the predominantly 2p-local bonding orbitals.

The grouping of valence orbitals predicted for Ce3 and Ce3O (Fig. 9) suggests that band X in both spectra can be assigned to detachment of an electron from their respective 6s-6s antibonding HOMO, band A in both spectra may be overlapping lower-intensity transitions associated with detachment from the three close-lying 5d-based bonding orbitals found ca. 0.5 eV lower in energy than the HOMO’s, and band B in both spectra may be associated with detachment of the in-phase 6s bonding orbitals predicted to be ca. 1 eV lower in energy than the HOMO’s. The difference in the relative energies of the 5d-based bonding orbitals for Ce3 and Ce3O relative to their HOMO’s, −0.7 eV and −0.3 eV, respectively, agrees with band A observed at lower eBE in the PE spectrum of Ce3O. Furthermore, the disparity between the β parameters of band A versus band X in both spectra might be expected for transitions associated with orbitals having significant differences in node structure.

The general decrease in higher eBE signal with increasing oxidation (Ce3O2, Ce3O3) is consistent with a depletion of electrons occupying the close-lying 5d and 6s-based MOs as well a bonding environment that further breaks the degeneracy of 5d and 6s-based molecular orbitals. With multiple electrons occupying numerous nearly degenerate molecular orbitals, the likelihood of short-lived electronic excitation of the anion followed by autodetachment to multiple close-lying neutral states is more likely. Additionally, shake-up (two electron) transitions may be occurring, as suggested for the smaller xy lanthanide oxide species. As oxidation increases, fewer electrons occupy any close-lying Ce-local orbitals that remain. The Ce3O4 PE spectrum, with a single dominant transition and clean baseline to higher binding energies, is easily reconciled with the calculations assuming that the spin-polarized orbitals shown in Fig. 11 correspond to a real picture of a doubly occupied 6s-based orbital. There are no other occupied orbitals other than the 4f core-like and 2p-local bonding orbitals. The spin of the remaining electron in the diffuse HOMO skimming the perimeter of the cluster is not coupled to the spin of the 4f-local electrons; transitions to either the higher or lower spin state are too close in energy to resolve: A comparable 4f 6sJa = 2 − Ja = 3 splitting in CeO is 82 cm−1,34 and this splitting could not be resolved in the PE spectrum of CeO measured with the same apparatus.50 

The PE spectra of Ce4Oy (y = 0–2) shown in Figure 13 exhibit intense, broad, and possibly overlapping transitions dominating the spectra at binding energies similar to those observed in spectra of smaller clusters. The EA’s of Ce4, Ce4O, and Ce4O2 are 0.97(5) eV, 0.94(2) eV, and 1.01(4) eV, respectively. As with the most reduced members of the Ce3Oy series, continuum signal is observed along with several distinct excited state transitions. Band X in the Ce4 spectrum is broad relative to Ce3, and the X-A energy interval, 0.8 eV, is larger than the X-A interval in Ce3. The spectra of Ce4O and Ce4O2 both exhibit a low-lying excited state transition, labeled A, overlapping with, and more intense than band X. Peak positions and asymmetry parameters are summarized in Table V.

TABLE V.

Summary of peak positions in the PE spectra of Ce4Oy and Ce5Oy shown in Figs. 13 and 14. The values in parentheses represent the uncertainty in the last digit.

ADE/VDE eBE/eVAsymmetry parameter
Ce4   
0.97(5)/1.10(1) 0.6(2) 
1.81(5)/1.85(3) 0.0(2) 
Ce4O   
0.94(2)/1.04(1) 0.6(2) 
VDE = 1.09(1) 0.5(2) 
VDE = 1.37(4) 0.2(2) 
Ce4O2   
1.01(4)/1.10(2) 0.5(1) 
VDE = 1.23(3) 0.4(1) 
Ce5O   
1.09(5)/1.22(1) 0.5(1) 
VDE = 1.34(3) 0.3(2) 
Ce5O2   
0.96(3)/1.15(2) 0.5(1) 
VDE = 1.35(3) 0.5(1) 
ADE/VDE eBE/eVAsymmetry parameter
Ce4   
0.97(5)/1.10(1) 0.6(2) 
1.81(5)/1.85(3) 0.0(2) 
Ce4O   
0.94(2)/1.04(1) 0.6(2) 
VDE = 1.09(1) 0.5(2) 
VDE = 1.37(4) 0.2(2) 
Ce4O2   
1.01(4)/1.10(2) 0.5(1) 
VDE = 1.23(3) 0.4(1) 
Ce5O   
1.09(5)/1.22(1) 0.5(1) 
VDE = 1.34(3) 0.3(2) 
Ce5O2   
0.96(3)/1.15(2) 0.5(1) 
VDE = 1.35(3) 0.5(1) 

The PE spectra of Ce5Oy (y = 1–2) shown in Figure 14 are also dominated by two comparably intense bands, X and A, and the continuum signal in both spectra is more intense than in the smaller clusters. The adiabatic electron affinities of Ce5O and Ce5O2 are 1.09(5) and 0.96(3) eV, respectively, comparable to all other clusters included in this study. Peak positions and asymmetry parameters are included in Table V.

Given the electronic structures predicted for the Ce2Oy and Ce3Oy cluster anions and neutrals, the positions and intensities of the Ce4Oy and Ce5Oy are consistent with detachment of electrons from molecular orbitals dominated by Ce 6s orbital contributions. The continuum signal will be considered more carefully in the discussion section.

The evolution of the electronic structure of these small cerium oxide clusters in the initial stages of oxidation provides a point of comparison to the bulk. Starting with the metallic species, the electronic orbital configuration of the Ce atomic 1G4 ground state is 4f 5d 6s2,57 which is not favorable for bonding67 The bulk band structure of metallic Ce decomposed into atomic orbitals has the 4f 5d2 6s occupancy, which is comparable to the orbital occupancy of Ce3, predicted by the DFT results. Likewise, Ce2, the spectrum of which could not be obtained, has analogous electronic structure based on high level ab initio calculations reported by Nokolaev,67 and Cao and Dolg.68 

Figure 15 shows a schematic of how the predicted electronic structure of the Ce3Oy series evolves with y, and how it compares the Ce2O3 bulk band structure based on DFT calculations reported by others.69,70 All Ce centers in the Ce3Oy clusters included in this study and bulk Ce2O3 have singly occupied 4f orbitals, which are situated at zero energy in the plot (EEf for bulk Ce2O3). The schematic shows that while the 6s- and 5d-based orbitals in Ce3 are interleaved energetically, electrons are depleted from the Ce-local 5d orbitals upon bonding with the O-atoms; the molecular orbitals with greatest Ce 5d contribution become unoccupied antibonding orbitals, which ultimately correlate to the conduction band of Ce2O3 and CeO2. The point at which x = y, each Ce atom has lost two electrons from the 5d orbitals, leaving each Ce center nominally with the 4f6s superconfiguration. While the companion to this report will show how subsequent oxidation shifts towards more distinctly ionic bonding, up to x = y, the Ce—O bonding can be described as largely ionic bonding, with some covalent character, as evidenced by the small contribution of Ce 5d orbitals in the 2p dominated bonding orbitals.

FIG. 15.

Schematic of electronic structure of the Ce3Oy series (y = 0–4) compared to the Ce2O3 bulk band structure69,70 based on DFT calculations.

FIG. 15.

Schematic of electronic structure of the Ce3Oy series (y = 0–4) compared to the Ce2O3 bulk band structure69,70 based on DFT calculations.

Close modal

The structures calculated for Ce3O4 and Ce3O4 are the same as the structure determined in IR predissociation measurements on Ce3O4+.44 This is a case in which a high-symmetry structure is unaffected by charge state, unlike several examples presented above, and the numerous cases we have found previously in transition metal suboxide anion and neutral structures.19 In accord with Asmis and co-workers,44 we note that the structures of Ce3O4, Ce3O4, and Ce3O4+ are evocative of the CeO2 (111) surface. The (111) surface of CeO2 was calculated to be the lowest energy surface,71 and though no detailed description of the Ce 4f orbital occupancy at the (111) surface, either O- or Ce-terminated in oxidizing or reducing environments has been published to our knowledge, the apparent stability of this common Ce3O4+/Ce3O4/Ce3O4 cluster may point to 4f orbital occupancy as a source of this stability. Oxygen termination of the reduced (111) surface has been proposed as a precursor to the phase transition between CeO2 and Ce2O3, for which there is the obvious change in 4f orbital occupancy.71 

The PE spectra of the range of suboxide clusters shown in this report have several peculiarities worth noting. First, the less oxidized the cluster, the more pronounced the continuum signal is above the adiabatic detachment energy. In addition to the occupied and partially occupied 5d-bonding orbitals, there are additionally low-lying 5d-based unoccupied orbitals (each Ce center has five 5d-orbitals, and only a maximum of 2 e per Ce atom are occupying them) so the anions and neutrals necessarily have numerous excited states associated with promoting the 5d-orbital electrons into other easily accessible 5d-orbitals. The continuum signal may conceivably be thermionic emission or autodetachment that follows electronic excitation. The fact that the continuum signal varies with photon energy and generally increases with cluster size (and therefore an increase in close-lying states) is consistent with the either picture, since both would be enhanced with increasing densities of electronic states.

Second, there are no significant variations in the electron affinities of the series of CexOy(y = 0–4) clusters included in this study. They all fall within several tenths of an eV from 1 eV, and based on the computational results, the ground state transitions are all associated with detachment of an electron from a delocalized molecular orbital that primarily has contributions from Ce 6s orbitals. This includes previously published results on Ce,72 CeO, Ce(OH)2,50 and unpublished result on larger CexOyHz clusters generated from CexOy + H2O reactions.59 Regardless of whether the 6s-based molecular orbitals are bonding or antibonding in character, the anisotropy parameters measured for the ground state transitions are all positive, with β(E) ∼ 0.5–1.5. This finding gives an interesting picture of how the electronic structure of a finite cerium suboxide cluster deviates from the bulk stoichiometric oxides, in which the 6s band lies high in energy above the 4f-5d conduction band: The cluster HOMO orbitals which are dominated by Ce-6s atomic orbitals are fairly non-bonding and delocalized over the whole cluster, and the detachment energy associated with these orbitals seems unaffected by size and extent of oxidation in the range of clusters we have studied. By extension, considering the propensity of metal oxides to accumulate negative charge under a range of applications, small cerium suboxide particles may have the metallic property of electron delocalization over the surface.

The anion photoelectron (PE) spectra of Ce2Oy(y = 1, 2), Ce3Oy(y = 0–4), Ce4Oy(y = 0–2), and Ce5Oy(y = 1, 2) were reported and analyzed with a qualitative insight into the cluster electronic structures from density functional theory calculations. The PE spectra all exhibit an intense electronic transition to the neutral ground state, all falling in the range of 0.7–1.1 eV electron binding energy, with polarization dependence consistent with detachment from Ce 6s-based molecular orbitals. There is no monotonic increase in electron affinity with increasing oxidation, which is direct evidence that the anions and neutrals are equally stabilized with stepwise oxidation.

A picture of how electronic structure evolves with oxidation state was formulated by comparing the experimental anion PE spectra and the computational results. The Ce2 metallic dimer was not observed experimentally, but the trends inferred for sequential oxidation of Ce2O to Ce2O2 were borne out in the more complete progression of oxidation determined for Ce3Oy. The electronic structure of the smallest metallic cluster observed in this study, Ce3, is consistent with the bulk structure in terms of atomic orbital occupancy (4f 5d2 6s). Initial cerium cluster oxidation involves largely ionic bond formation via Ce 5d and O 2p orbital overlap (i.e., O 2p localized), with Ce—O—Ce bridge bonding favored over Ce=O terminal bond formation. With subsequent oxidation, the Ce 5d-based molecular orbitals are depleted of electrons, with highest occupied orbitals described as diffuse Ce 6s based molecular orbitals. In the y ≤ (x + 1) range of oxidation states, each Ce center has a singly occupied non-bonding 4f orbital. The PE spectrum of Ce3O4 is unique in that it exhibits a single nearly vertical transition. The highly symmetric structure predicted computationally is the same structure determined from Ce3O4+ IR predissociation spectra,44 which resembles the Ce-terminated (111) surface of CeO2.

Spectra of clusters with x ≥ 3 exhibit considerable continuum signal above the ground state transition; the intensity of the continuum signal decreases with increasing oxidation. This feature was attributed to a plethora of partially occupied, nearly degenerate Ce 5d and 6s-based molecular orbitals, which give rise to the possibility of numerous close-lying quasibound anion states that undergo autodetachment to a range of neutral states, as well as two-electron transitions. Upon oxidation, the nearly degenerate orbitals separate energetically into O 2p-local bonding and Ce 5d-local anti-bonding orbitals while the O-atoms accumulate electrons from the previously partially occupied 5d and 6s-based orbitals, decreasing the density of low-lying electronic states in both the anion and neutral.

See supplementary material for details on spectroscopic parameters used in the CexOy PE spectral simulations, comparison of Ce2O two disparate 5d-based orbital configurations, diagrams summarizing various cluster structural isomers and their relative energies, and a complete list of Cartesian coordinates for all structures that converged in the calculations.

This work was supported by the National Science Foundation Grant No. CHE-1265991.

1.
T.
Nakazawa
,
T.
Inoue
,
M.
Satoh
, and
Y.
Yamamoto
,
Jpn. J. Appl. Phys., Part 1
34
,
548
(
1995
).
2.
N. V.
Skorodumova
,
R.
Ahuja
,
S. I.
Simak
,
I. A.
Abrikosov
,
B.
Johansson
, and
B. I.
Lundqvist
,
Phys. Rev. B
64
,
115108
(
2001
).
3.
Z.
Shao
,
S. M.
Haile
,
J.
Ahn
,
P. D.
Ronney
,
Z.
Zhan
, and
S. A.
Barnett
,
Nature
435
,
795
798
(
2005
).
4.
D. R.
Mullins
,
Surf. Sci. Rep.
70
,
42
(
2015
).
5.
R. J.
Gorte
,
AIChE J.
56
,
1126
(
2010
).
6.
A.
Bruix
,
J. A.
Rodriguez
,
P. J.
Ramírez
,
S. D.
Senanayake
,
J.
Evans
,
J. B.
Park
,
D.
Stacchiola
,
P.
Liu
,
J.
Hrbek
, and
F.
Illas
,
J. Am. Chem. Soc.
134
,
8968
(
2012
).
7.
E. S.
Bickford
,
S.
Velu
, and
C.
Song
,
Catal. Today
99
,
347
(
2005
).
8.
T.
Bunluesin
,
R. J.
Gorte
, and
G. W.
Graham
,
Appl. Catal., B
15
,
107
(
1998
).
9.
A.
Luengnaruemitchai
,
S.
Osuwan
, and
E.
Gulari
,
Catal. Commun.
4
,
215
(
2003
).
10.
Y.
Li
,
Q.
Fu
, and
M.
Flytzani-Stephanopoulos
,
Appl. Catal., B
27
,
179
(
2000
).
11.
M. V.
Ganduglia-Pirovano
,
A.
Hofmann
, and
J.
Sauer
,
Surf. Sci. Rep.
62
,
219
(
2007
).
12.
13.
N. V.
Skorodumova
,
S. I.
Simak
,
B. I.
Lundqvist
,
I. A.
Abrikosov
, and
B.
Johansson
,
Phys. Rev. Lett.
89
,
166601
(
2002
).
14.
M. S.
Dresselhaus
and
I. L.
Thomas
,
Nature
414
,
332
(
2001
).
15.
J.
Vecchietti
,
A.
Bonivardi
,
W.
Xu
,
D.
Stacchiola
,
J. J.
Delgado
,
M.
Calatayud
, and
S. E.
Collins
,
ACS Catal.
4
,
2088
(
2014
).
16.
G.
Pacchioni
,
Phys. Chem. Chem. Phys.
15
,
1737
(
2013
).
17.
J.
Paier
,
C.
Penschke
, and
J.
Sauer
,
Chem. Rev.
113
,
3949
(
2013
).
18.
J.
Carrasco
,
D.
López-Durán
,
Z.
Liu
,
T.
Duchon̆
,
J.
Evans
,
S. D.
Senanayake
,
E. J.
Crumlin
,
V.
Matolin
,
J. A.
Rodriguez
, and
M. V.
Ganduglia-Pirovano
,
Angew. Chem., Int. Ed.
54
,
3917
(
2015
).
19.
J. E.
Mann
,
N. J.
Mayhall
, and
C. C.
Jarrold
,
Chem. Phys. Lett.
525
,
525
526
(
2012
).
20.
D. W.
Rothgeb
,
J. E.
Mann
, and
C. C.
Jarrold
,
J. Chem. Phys.
133
,
054305
(
2010
).
21.
M.
Ray
,
S. E.
Waller
,
A.
Saha
,
K.
Raghavachari
, and
C. C.
Jarrold
,
J. Chem. Phys.
141
,
104310
(
2014
).
22.
23.
J. A.
Rodriguez
,
P.
Liu
,
J.
Hrbek
,
J.
Evans
, and
M.
Pérez
,
Angew. Chem. Int. Ed.
46
,
1329
(
2007
).
24.
H.
Wu
,
N.
Zhang
,
H.
Wang
, and
S.
Hong
,
Chem. Phys. Lett.
568
,
84
(
2013
).
25.
S. M.
Lang
and
T. M.
Bernhardt
,
Phys. Chem. Chem. Phys.
14
,
9255
(
2012
).
26.
J. A.
Felton
,
M.
Ray
,
S. E.
Waller
,
J. O.
Kafader
, and
C. C.
Jarrold
,
J. Phys. Chem. A
118
,
9960
(
2014
).
27.
X. N.
Wu
,
X. L.
Ding
,
S. M.
Bai
,
B.
Xu
,
S. G.
He
, and
Q.
Shi
,
J. Phys. Chem. C
115
,
13329
(
2011
).
28.
F.
Aubriet
,
J. J.
Gaumet
,
W. A.
de Jong
,
G. S.
Groenewold
,
A. K.
Gianotto
,
M. E.
McIlwain
,
M. J.
Van Stipdonk
, and
C. M.
Leavitt
,
J. Phys. Chem. A
113
,
6239
(
2009
).
29.
S.
Hirabayashi
and
M.
Ichihashi
,
J. Phys. Chem. A
117
,
9005
(
2013
).
30.
S.
Hirabayashi
and
M.
Ichihashi
,
Chem. Phys. Lett.
564
,
16
(
2013
).
31.
T.
Nagata
,
K.
Miyajima
,
R. A.
Hardy
,
G. F.
Metha
, and
F.
Mafuné
,
J. Phys. Chem. A
119
,
5545
(
2015
).
32.
T.
Nagata
,
K.
Miyajima
, and
F.
Mafuné
,
J. Phys. Chem. A
119
,
10255
(
2015
).
33.
T.
Nagata
,
K.
Miyajima
, and
F.
Mafuné
,
J. Phys. Chem. A
119
,
1813
(
2015
).
34.
L. A.
Kaledin
,
J. E.
McCord
, and
M. C.
Heaven
,
J. Mol. Spectrosc.
158
,
40
(
1993
).
35.
L. A.
Kaledin
,
J. E.
McCord
, and
M. C.
Heaven
,
J. Mol. Spectrosc.
170
,
166
(
1995
).
36.
C.
Linton
,
J.
Chen
, and
T. C.
Steimle
,
J. Phys. Chem. A
113
,
13379
(
2009
).
37.
C.
Linton
,
M.
Dulick
,
R. W.
Field
,
P.
Carette
,
P. C.
Leyland
, and
R. F.
Barrow
,
J. Mol. Spectrosc.
102
,
441
(
1983
).
38.
R. F.
Barrow
,
R. M.
Clemments
,
S. M.
Harris
, and
P. P.
Jenson
,
J. Astrophys.
229
,
439
(
1979
).
39.
H.
Moriyanma
,
H.
Tatewaki
, and
S.
Yamamoto
,
J. Chem. Phys.
138
,
224310
(
2013
).
40.
M.
Dolg
,
H.
Stoll
, and
H.
Preuss
,
J. Mol. Struct.: THEOCHEM
231
,
243
(
1991
).
41.
T. K.
Todorova
,
I.
Invfante
,
L.
Gagliardi
, and
J. M.
Dyke
,
Int. J. Quantum. Chem.
109
,
2068
(
2009
).
42.
M.
Dolg
, “
Lanthanides and actinides
,” in
Encyclopedia of Computational Chemistry
(
John Wiley & Sons, Ltd.
,
2002
), Published Online: 15 April 2002.
43.
R. W.
Field
,
Ber. Bunsengesellschaft Phys. Chem.
86
,
771
(
1982
).
44.
A. M.
Burow
,
T.
Wende
,
M.
Sierka
,
R.
Wlodarczyk
,
J.
Sauer
,
P.
Claes
,
L.
Jiang
,
G.
Meijer
,
P.
Lievens
, and
K. R.
Asmis
,
Phys. Chem. Chem. Phys.
13
,
19393
(
2011
).
45.
L.
Jiang
,
T.
Wende
,
P.
Claes
,
S.
Bhattacharyya
,
M.
Sierka
,
G.
Meijer
,
P.
Lievens
,
J.
Sauer
, and
K. R.
Asmis
,
J. Phys. Chem. A
115
,
11187
(
2011
).
46.
S. P.
Willson
and
L.
Andrews
,
J. Phys. Chem. A
103
,
3171
(
1999
).
47.
S. P.
Willson
and
L.
Andrews
,
J. Phys. Chem. A
102
,
10238
(
1998
).
48.
S. P.
Willson
and
L.
Andrews
,
J. Phys. Chem. A
104
,
1640
(
2000
).
49.
S. P.
Willson
,
L.
Andrews
, and
M.
Neurock
,
J. Phys. Chem. A
104
,
3446
(
2000
).
50.
M.
Ray
,
J.
Felton
,
J. O.
Kafader
,
J.
Topolski
, and
C. C.
Jarrold
,
J. Chem. Phys.
142
,
064305
(
2014
).
51.
C.-X.
Chi
,
H.
Xie
,
R.
Cong
,
Z.-C.
Tang
, and
M.-F.
Zhou
,
Chin. J. Chem. Phys.
24
,
604
(
2011
).
52.
R.
Klingeler
,
G.
Lüttgens
,
N.
Pontius
,
R.
Rochow
,
P. S.
Bechtold
,
M.
Neeb
, and
W.
Eberhardt
,
Eur. Phys. J. D
9
,
263
(
1999
).
53.
R.
Klingeler
,
N.
Pontius
,
G.
Lüttgens
,
P. S.
Bechthold
,
M.
Neeb
, and
W.
Eberhardt
,
Phys. Rev. A
65
,
032502
(
2002
).
54.
A. S.
Ivanov
,
X.
Zhang
,
H.
Wang
,
A. I.
Boldyrev
,
G.
Gantefoer
,
K. H.
Bowen
, and
I.
Cernusak
,
J. Phys. Chem. A
119
,
11293
(
2016
).
55.
J. O.
Kafader
,
M.
Ray
, and
C. C.
Jarrold
,
J. Chem. Phys.
142
,
034305
(
2015
).
56.
J. O.
Kafader
,
M.
Ray
, and
C. C.
Jarrold
,
J. Chem. Phys.
143
,
064305
(
2015
).
57.
A.
Kramida
,
Yu.
Ralchenko
,
J.
Reader
, and
NIST ASD Team
, NIST Atomic Spectra Database (version 5.3), National Institute of Standards and Technology, Gaithersburg, MD, 2015, available: http://physics.nist.gov/asd, 8 April 2016.
58.
S. E.
Waller
,
M.
Ray
,
B. L.
Yoder
, and
C. C.
Jarrold
,
J. Phys. Chem. A
117
,
13919
(
2013
).
59.
J. E.
Topolski
,
J. O.
Kafader
,
M.
Ray
, and
C. C.
Jarrold
, Elucidating cerium + H2 O reactivity through electronic structure: A combined PES and DFT study, (unpublished).
60.
V. D.
Moravec
and
C. C.
Jarrold
,
J. Chem. Phys.
108
,
1804
(
1998
).
61.
S. E.
Waller
,
J. E.
Mann
, and
C. C.
Jarrold
,
J. Phys. Chem. A
117
,
1765
(
2013
).
62.
M. J.
Frisch
,
G. W.
Trucks
,
H. B.
Schlegel
,
G. E.
Scuseria
,
M. A.
Robb
,
J. R.
Cheeseman
,
G.
Scalmani
,
V.
Barone
,
B.
Mennucci
,
G. A.
Petersson
,
H.
Nakatsuji
,
M.
Caricato
,
X.
Li
,
H. P.
Hratchian
,
A. F.
Izmaylov
,
J.
Bloino
,
G.
Zheng
,
J. L.
Sonnenberg
,
M.
Hada
,
M.
Ehara
,
K.
Toyota
,
R.
Fukuda
,
J.
Hasegawa
,
M.
Ishida
,
T.
Nakajima
,
Y.
Honda
,
O.
Kitao
,
H.
Nakai
,
T.
Vreven
,
J. A.
Montgomery
, Jr.
,
J. E.
Peralta
,
F.
Ogliaro
,
M. J.
Bearpark
,
J.
Heyd
,
E. N.
Brothers
,
K. N.
Kudin
,
V. N.
Staroverov
,
R.
Kobayashi
,
J.
Normand
,
K.
Raghavachari
,
A. P.
Rendell
,
J. C.
Burant
,
S. S.
Iyengar
,
J.
Tomasi
,
M.
Cossi
,
N.
Rega
,
N. J.
Millam
,
M.
Klene
,
J. E.
Knox
,
J. B.
Cross
,
V.
Bakken
,
C.
Adamo
,
J.
Jaramillo
,
R.
Gomperts
,
R. E.
Stratmann
,
O.
Yazyev
,
A. J.
Austin
,
R.
Cammi
,
C.
Pomelli
,
J. W.
Ochterski
,
R. L.
Martin
,
K.
Morokuma
,
V. G.
Zakrzewski
,
G. A.
Voth
,
P.
Salvador
,
J. J.
Dannenberg
,
S.
Dapprich
,
A. D.
Daniels
,
Ö.
Farkas
,
J. B.
Foresman
,
J. V.
Ortiz
,
J.
Cioslowski
, and
D. J.
Fox
, Gaussian 09, Gaussian, Inc., Wallingford, CT, USA, 2009.
63.
X.
Cao
and
M.
Dolg
,
J. Chem. Phys.
115
,
7348
(
2001
).
64.
S. M.
O’Malley
and
D. R.
Beck
,
Phys. Rev. A
74
,
042509
(
2006
).
65.
B.
Johansson
,
W.
Luo
,
S.
Li
, and
R.
Ahuja
,
Sci. Rep.
4
,
6398
(
2014
).
66.
B. L.
Yoder
,
J. T.
Maze
,
K.
Raghavachari
, and
C. C.
Jarrold
,
J. Chem. Phys
122
,
094313
(
2005
).
67.
A. V.
Nikolaev
,
Phys. Rev. A
84
,
012512
(
2011
).
68.
X.
Cao
and
M.
Dolg
,
Mol. Phys.
101
,
1967
(
2003
).
69.
J. L. F.
Da Silva
,
M. V.
Ganduglia-Pirovano
,
J.
Sauer
,
V.
Bayer
, and
G.
Kresse
,
Phys. Rev. B
75
,
045121
(
2007
).
70.
P. J.
Hay
,
R. L.
Martin
,
J.
Uddin
, and
G. E.
Scuseria
,
J. Chem. Phys.
125
,
034712
(
2006
).
71.
M.
Fronzi
,
A.
Soon
,
B.
Delley
,
E.
Traversa
, and
C.
Stampfl
,
J. Chem. Phys.
131
,
104701
(
2009
).
72.
J.
Felton
,
M.
Ray
, and
C. C.
Jarrold
,
Phys. Rev. A
89
,
033407
(
2014
).

Supplementary Material